Automatic pattern extraction from polyphonic MIDI files. (Extraction automatique de patterns à partir de séquences polyphoniques MIDI.) - Institut de Recherche et Coordination Acoustique/Musique Accéder directement au contenu
Communication Dans Un Congrès Année : 2003

Automatic pattern extraction from polyphonic MIDI files. (Extraction automatique de patterns à partir de séquences polyphoniques MIDI.)

Résumé

In the context of musical analysis, we propose an algorithm that automatically induces patterns from polyphonies. We define patterns as "perceptible repetitions in a musical piece". We claim that a link can be established between the patterns that we extract and the musical structure of the piece. Our aim is that the patterns we extract are perceptively relevant although several perceptive criteria are not taken into account in our model: what we propose is an attempt to explore the limits of a system that do not considers, in a first step, the musical notions of expectation or temporal context (considering temporal context means that we take into account past events for the analysis of current events), but that integrates several other perceptive notions such as polyphonic context. This brings us to discuss several specific issues related to the extraction of patterns in a polyphonic context. Our model proceeds as follows: In a first step, we quantize a MIDI sequence with an algorithm described in [1], and we segment the music in "beat segments". Then, we compute a similarity matrix from the segmented sequence. The measure of similarity we use is described in [2]. The algorithm computes the similarity between all the pairs of sequences with same number of beats. It relies on features such as rhythm, contour and pitch intervals. Last, a bottom-up approach is proposed for extracting patterns from the similarity matrix. It builds new patterns by concatenation of the consecutive cells that are on the same diagonal in the similarity matrix. The algorithm was tested on several pieces of music, and interesting results were found. At the same time, new questions were raised on the notion of musical structure.
Fichier principal
Vignette du fichier
meudic.pdf (2.63 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02994194 , version 1 (07-11-2020)

Identifiants

  • HAL Id : hal-02994194 , version 1

Citer

Benoit Meudic. Automatic pattern extraction from polyphonic MIDI files. (Extraction automatique de patterns à partir de séquences polyphoniques MIDI.). Journées d'Informatique Musicale, Jun 2003, Montbéliard, France. ⟨hal-02994194⟩

Collections

IRCAM AFIM JIM
55 Consultations
83 Téléchargements

Partager

Gmail Facebook X LinkedIn More